Advancing CPR Analysis: Ear ECG and Two-Stage Data-Driven Adaptive Filtering

This study investigates ear ECG (eECG) as an alternative to traditional chest ECG during cardiopulmonary resuscitation (CPR) to maintain continuous chest compressions. We present a novel data-driven adaptive forgetting factor sparse recursive least squares (D2AF-SRLS) method for artifact removal, implemented in two stages, which is both adaptable and computationally efficient. A novel, non-invasive PDMS-based eECG electrode manufactured in a low-cost, scalable technique is meant to improve ear canal contact and sensing. The performance of six rat subjects is evaluated after 7 minutes of cardiac arrest, followed by manual chest compressions for CPR, and the results of the eECG are compared to chest ECG recordings. To our knowledge, this is the first study to use a PDMS eECG electrode with a data-adaptable filtering approach. The D2AF-SRLS algorithm achieved a mean SNR improvement of 20 dB, exceeding state-of-the-art approaches. Our approach provides proof of concept for the eECG technique and artifact-free recording during CPR. Our work lays the groundwork for translation to uninterrupted clinical monitoring of patients. This eliminates the necessity to pause for rhythm analysis, potentially enhancing the efficiency and effectiveness of CPR procedures.